GMTK Game Jam 2020

This week I was able to participate in the 2020 Game Makers Toolkit Game Jam and made the game Oliver Plays The Labyrinth. I've been a fan of Mark Brown and his channel for years but this was the first chance I had to participate in one of his jams. In fact it was the first jam I've participated in for several years. It's certainly not as easy to clear a whole weekend as it once was. I would say I was able to jam for around 25 of the given 48 hours, but I'm pretty happy with the outcome and this jam reminded me of how valuable jamming can be as creative exercise. I certainly hope to do them more regularly going forward.

Play the game at

The theme of this years jam was “out of control”. Since I have primarily been working on VR games the last few years my first inclination was to make one for the jam. However, the Jam's first round of scoring is done by other jammers, and it's unlikely many would be able to try a game that requires a headset.

My thought then was "how do I make a VR game that doesn't require VR?". I thought that being taken "out of control" of the actual character of the game and instead being in control of the person who is in control of the main character could be interesting.

So I began thinking of different ways you could control someone who was in VR. My first few ideas were too complicated, along the lines of convoluted QWOP style hand manipulation. But the nature of the jam forced me to quickly pare down ideas to only what was necessary.

Room scale movement seemed like the most obvious element of VR to focus on. I decided that having to navigate around obstacles in VR and non-VR simultaneously was an interesting challenge given this setup and I used a top-down orthographic camera for the non-VR world to contrast the two perspectives as much as possible.

I put together a few puzzle layouts:
This browser does not suppoer webP images This browser does not suppoer webP images This browser does not suppoer webP images

I sourced some free-to-use 2d and 3D art assets from Pinterest and the Unity asset store. Oliver himself I mostly made myself though I used some free-to-use sprites as a working base. What do I think worked? I think the concept is sound; having to mentally project the two perspectives onto each other to move around is interesting to me.

I also think it's highly scalable. There are many more mechanics I have in mind to introduce to the VR world that throws problems at the player. And navigating the non-VR world will be hilarious when Oliver has to leave his room and move throughout the house or even outside.

What doesn't work? The controls are very awkward when you are looking at the VR perspective. Of course this is more or less intentional and baked into the concept, but I'm not sure it supports longer play sessions. I think adding separate controls that allow you to instead move according to the VR perspective could hold the complexity but add some comfort. We'll see. The art is fine by solo non-artist jam standards but I would love to see improvements to both the VR and non-VR art.

I would also love to see Oliver play other game sin the future. Having him play a shooter or an rpg would be especially interesting. Also adding some tasks to the non-VR world could be fun. Can Oliver beat The Labyrinth while taking out the trash?

Economizing Verbs in VR

In traditional game design it is generally considered best to be as economical as possible with your verbs. Don't use two when one would do. For example in Grand Theft Auto hijacking a car, breaking into a car, and getting into an unlocked car all require the same input and whether you are using a key to start a car or having to hot-wire it these are done automatically. In the Legend of Zelda series players can break pots, but there is no unique verb for doing so. Players break pots by using several other multi-purpose verbs - swinging their sword, rolling, bombing, picking up and throwing, shooting with an arrow, etc.

It's easy to imagine all of these interactions having more unique verbs but unless adding them would achieve a specific design goal they may make gameplay feel more cumbersome as verbs players will have to perform frequently must be very rewarding to not become monotonous. Sometimes the easiest way to make a verb feel good is to minimize the amount of thinking a player has to do to perform it, or to remove it's input entirely and make it passive (such as with hot-wiring cars in GTA ).

On the other hand when designers add verbs players rarely perform they risk creating orphaned verbs. These are verbs which have no relationship to the other verbs in the game. Orphaned verbs can be especially troublesome if are rarely used. Sometimes it takes a google search to remember that the feature exists at all because you learned it 10 hours ago in a tutorial but now you're stuck without it.

Parrying seems to be a feature that often becomes an orphaned verb. Games such as Dark Souls and Breathe of the Wild feature parrying but do not require players to use it. However both games feature enemy encounters near the end of the game which are difficult if you have not practiced parrying. On message boards you find players struggling with these late game enemies because they never practiced parrying or completely forgot about it.

However, the purpose of this post isn't to talk about standard game verb concepts, it is to talk about how virtual reality is overturning some of these conventional design rules. VR has changed input from something that has to be learned into something often instinctual, and it's completely changed what players expect from game verbs.

In VR the abstraction from input to verb can be greatly reduced. Every aspect of video game interaction has become subject to developers swapping button presses for physical movements. Picking up an object no longer has to be mapped to using a joystick to walk into it, you can physically reach out to grab the object.

Because of this immersion, instead of players assuming verbs will be mapped to the same input that a completely separate verbs are (i.e. how in non-VR games opening a door and opening a chest are often done with the same button) players will often expect each verb to have a unique input that to some degree simulates it's real life equivalent.

I ran a small experiment once with the original Oculus touch controllers in which I put friends into a VR experience in which they stood in a realistic looking kitchen surrounded by ordinary kitchen supplies. I asked them just to interact with objects around them. The trick was that I had programmed large items, such as pots and pans, to be picked up using the grip button while smaller items, like forks and spoons, required the trigger.

The grip button is the button most commonly used for grabbing in VR, and the motion players make to press it most closely simulates a grab. The trigger button, however, makes players use their index finger and simulate a pinch.

Without receiving any instructions all of my friends instinctively used the correct buttons when picking up the differently sized objects. After they had done so I asked them if they noticed a difference in how they picked up each item. They did not.

These instinctual interactions can be seen all the time in VR. Often players expect there to be verbs where there are none. Players try to pop open soda cans, try to push buttons on appliances, put their hands in front of other characters' faces to see if they will react realistically, etc.

The challenge in VR is not necessarily to economize verbs, but to balance the amount and complexity of verbs while accommodating both for players' instinctual assumptions and their desire for ease of use. Players don't want VR games in which stealing a car feels the same as getting in your own car and they don't want cars to hotwire themselves (in part because your avatar acting autonomously doesn't make sense in VR, but that's a different topic), yet they also don't want these verbs to be as complex and involved as they are in real life. Players still want verbs abstracted enough so they are fun.

Of course taken to it's natural extreme extracting these "instinctual" verbs from player expectation can create an unreasonable amount of interactions. I once had a client who was unfamiliar with VR (and video games in general) report a bug because the laptop in the game "didn't work" - as in she wasn't able to use it like a real computer and surf the internet. That's when I realized the power of virtual reality to delight but also disappoint. So where do we draw the line?

We see some games effectively push the envelope on the complexity of verbs and interaction. The game Onward has weapons with semi-realistic reloading mechanics. Many guns have unique ways to reload them so if you start a match with a gun you haven't used before there's a chance you'll be asking your teammates how to reload while taking cover from an incoming spray of bullets. This works great for a mil-sim like Onward but is probably much too complex to be a "best-practice" type approach for the industry.

Half Life Alyx introduces a very clever way to make focus the complexity of their reloading mechanic. When the player reaches over their shoulder to grab something out of their backpack they will automatically grab whatever it is they were looking for. It's kind of magical. Even without having to specify what you want to pull out of your inventory you instantly find whatever ammo the gun you're holding uses, or if your weapon is in an upgrade chamber you grab the resource needed to complete the upgrade. I would describe this as a contextual verb, something that is a staple of non-VR games but less common in VR, and is a smart way to accommodate for the expectation of physical interactions in VR while not increasing the complexity of a verb you want to remain simple.

The rest of Alyx's reload mechanic is, by design, much more complicated as players fumble to insert clips or shells before advancing head crabs suck their face off. On the other hand Valve handles switching weapons with a menu that players navigate by hand movement. It's interesting to see one game utilizing so many different verbs of varying levels of complexity and abstraction. Each one designed perfectly to put the tension and anxiety where Valve wanted it to be.

In Death, a good VR rogue-like with an excellent bow and arrow mechanic, half-economizes their movement verb by giving players 2 different methods of teleportation. One is a unique verb that requires the player to throw a shard to the spot they want to teleport, and the second recycles the bow and arrow mechanic by allowing the player to shoot an arrow to where they want to teleport.

The first method (which by the way is fantastic in my opinion) is a unique verb as grabbing and throwing is not done at all for any other reason in the game. I have seen this mechanic become orphaned, when watching friends play, as it can be easily forgotten versus the second method which is seemingly more apparent since players are already constantly using the bow and arrow.

Pistol Whip economizes verbs very cleanly by allowing the player to do almost everything with a bullet. Choosing a level, customizing your gun, changing settings, it's all done by shooting something with your gun, the same verb that is used for 90% of the gameplay.

At this point it seems that there is no one right way to economize verbs in VR. We see brilliance in economizing verbs as much as possible in Pistol Whip, while Alyx leans into a variety of verbs while also looking for ways to economize, while simulation games such as Onward thrive in their complexity of verbs. It seems that instead of seeing standard best-practices form in the near future we will see designers exploring and taking advantage of the breadth of possibilities VR offers.

Gunner Punks Dev Blog: Mounted Gun Part 1

One of the most unique aspect of Gunner Punks gameplay is the mounted gun. It has 4 axis of rotation which allows players to position it in any angle needed. There seems to be a consistent eureka moment for play testers when they realize the gun mechanically works exactly as you would expect in real life. It's one of those cool interactions that has only been made possible through VR.

Mechanically it uses IK for arm movement and a hinge joint for rotation. One of the early challenges with getting it to operate smoothly for the player was determining the right number of arms and the length of the arms.

Initially 2 arms was introducing some difficulty to gameplay. There were a number of positions it was difficult or impossible to maneuver the gun into. The rotation of the arms would compound to make certain movements faster than expected while others were slower, which lead to player over-compensating and in general thinking too much and having to “figure out” the weapons movements.

Ultimately shortening the length of the bottom arm fixed a lot of the awkwardness, but another helpful step was flipping it's direction as well. The bottom arm was originally pointing away from the player, or in the same direction as the barrel, which is actually more consistent with how this would be constructed in real life. You wouldn't want the arm coming down on and hitting you right in the chest.

But with that setup almost every movement required rotation from both arms. Flipping the bottom arm allows movement forwards and backwards to be done with the upper arms, while the lower arm almost exclusively handles vertical movement. This isolation creates much more predictable and fluid movements.

I also experimented with 1 and 3 arms but each introduced issues. 1 arm had a more fluid and consistent movement, but was very limited in it's maneuverability and forced the player to be very rigid in their body movement. In particular being able to pull the weapon to your chest when you wanted to shoot upward was no longer an option and would instead cause the gun to point downward. It was unnatural.

3 arms gave a great range of positions the gun could maneuver into but the awkwardness of so many arms introduced several positions in which they could be folded in a way that would bind them awkwardly if the player did not unfold them in the correct way. Limiting their rotation to prevent this again caused a rigidity that offset the advantage of the third arm.

Ultimately the movement has become very fluid and natural to handle and achieves allowing the player to quickly slip into a flow state without being jarred by unexpected behavior. In part 2 I will talk about some of the challenges faced deciding how balance simplicity vs presence in regards to the different ways of interacting with the gun.

The Addiction of Dissapointment

Do mainstream gamers enjoy extremely difficult games? For the last 20 years the industry consensus seemed to be that most do not. Instead mainstream games have been trending toward accessibility and positive reinforcement, especially for games aimed at younger audiences. Frequent save points, quest markers, low death rates, and general hand holding have become design staples. There has been an inverse to this trend, with genres such as souls-like and rogue-like gaining mainstream popularity and a certain "cool kid" appeal. But these are mostly seen as targeting "hardcore" and mature gamers, and are considered somewhat niche.

However, in 2017 a new genre emerged and quickly dominated the gaming world. Battle Royals attract hundreds of millions of gamers of all ages, all gaming backgrounds, and on every device. Yet Battle Royals are also one of the hardest genre's on the market. Few players will ever win a single game. Many, likely even most, players have a k/d ratio less than 1. In any given game only 1-4 players out over 100+ can be victorious. In spite of all of this battle royals have become the most popular genre for casual gamers. How did this happen? It seems that souls-likes, rogue-likes, and especially battle royals use failure and disappointment to create addictive game loops that keeps players coming back again and again.

Honestly I rarely walk away from playing battle royals feeling great. Regardless of how well I do they usually put me in a bad mood. Yet I keep playing, and when I do I have a hard time stopping. One more game. One more game. One more game.

One night playing Call of Duty Warzone I could only wonder 'why am I doing this to myself', and it occurred to me what a brilliant game loop I had fallen prey to. Almost every time I am considering whether to jump into another game it's while I am feeling disappointment. To illustrate this I wish I had an audio montage of all the conversations my squad has had in the menu after games. "We should have won that", "so close", "that was garbage", "I hate campers", "they must be hacking", "how did that happen", etc.

The magic is that while my squad is almost frustrated after every game, the general memory of the game is usually that it was fun. Why? Because a lot of good things happened in the middle of the game. Maybe we picked up great weapons, maybe we completed a contract, maybe we got a few kills. Enough went our way that we felt we had a chance, we had hope, and we had fun.

Games seem to follow the same pattern for most players. Optimism when they drop in, excitement when they are looting and finding weapons, a rush for every kill they get, and then eventually and almost inevitably disappointment when they die for reasons they are likely to declare were bs.

The key to creating hope in the middle of this loop is the randomness of each game. Because so many elements are randomized it's easy for players to believe that this game, or the next game, or the next game, will be different. Winning in some ways is just waiting for the right combination of of variables. It's a slot machine. Souls-like games do not have as much randomization, so while they have a similar emotional game loop the amount of optimism you feel is proportional to your faith in your skills, and the inevitable disappointment you experience can specifically work against this confidence.

Rogue-likes and rogue-lites have virtually the same loop as Battle Royals including the randomization, perma-death or semi-perma-death, and constant crushing defeat. So why aren't they as popular as battle royals? Clearly the multiplayer component is the difference maker here. Of course it adds a social component, but in terms of the game loop it adds so many additional layers to the randomization that AI doesn't. It at doubles or triple the number of slots on the slot machine.

Battle royals are like slot machine if the casino could guarantee that each player would consistently get good spins but they can only walk away from the machine after a bad spin or after hitting the jackpot. Of course this is exactly the behavior casinos want to create, but Battle Royals guarantee it.

It' worth noting how Battle Royals are also different than "traditional" FPS multiplayer death match. These games features a series of up and down through a number of kills, deaths, weapon pickups, weapons drops, etc. The end of the games are equal parts exciting or disappointing for team based play (which is most common) since half of the player win.

The gaming industry has always incorporated elements of gambling into their designs, and the last decade in particular saw many new attempts to replicate the level of addiction seen in casinos. Social games, loot boxes, and now battle royals carefully integrate tested practices that keep players hooked. In part 2 we will look at some of the studies of gambling behavior and see how science shows that disappointment truly is addictive.